Utility Representation for Collaborative Surveillance
Geoffrey Dolinger, Timothy Sharp, Alexander Stringer, Joseph Karch, Asad Vakil, Justin G. Metcalf, Adam Bowersox · 2025
Surveillance of physical space is a balance of exploration of unknown regions and exploitation of domain knowledge to detect targets of interest. The search function often illuminates all space with equal priority but when resources are limited and information regarding the environment is present then optimal search becomes increasingly important. This paper presents an approach to optimize collaborative search via a utility representation. This method is a command and control (C2) inspired search representation which incorporates domain information in the environment, the radar platform's location, and a beta distribution to balance new space exploration and returning to high priority locations. Testing in a space-time adaptive processing (STAP) radar simulation shows this method detects more targets and scales with multiple platforms. The spatial representation and action selection method automatically coordinate across large distances and can incorporate shared information for expanded collaboration.